Mediators of the effects on fatigue of pragmatic rehabilitation for chronic fatigue syndrome.

Abstract

To examine potential mediators of the effect of pragmatic rehabilitation on improvements in fatigue following a randomized controlled trial for patients with chronic fatigue syndrome (CFS/ME) in primary care (IRCTN 74156610). Patients fulfilled the Oxford criteria for CFS. Ninety-five patients were randomized to pragmatic rehabilitation and 100 to general practitioner (GP) treatment as usual. The outcome was the Chalder fatigue scale score (0123 scoring) at end of treatment (20 weeks) and 1-year follow up (70 weeks). First, the effect of treatment on potential mediators was assessed. Then fatigue was regressed on significant mediators, treatment allocation, and baseline measures of fatigue and significant mediators. Reduction in limiting activities at 20 weeks mediated the positive effect of pragmatic rehabilitation on fatigue at 70 weeks (mediated effect size = -2.64, SE = 0.81, p = .001, proportion of effect mediated = 82.0%). Reduction in catastrophizing at 20 weeks mediated the positive effect of pragmatic rehabilitation on fatigue at 70 weeks (mediated effect size = -1.39, SE = 0.61, p = .023, proportion of effect mediated = 43.2%). Reductions in 70-week measures of fear avoidance, embarrassment avoidance, limiting activities, and all-or-nothing behavior all mediated improvement in fatigue at 70 weeks, although the causal direction of these cross-sectional effects cannot be determined. There were no between-group differences on measures of exercise capacity (a timed step test). Improvements in fatigue following pragmatic rehabilitation are related to changes in behavioral responses to and beliefs about fatigue.